Therapeutic Interventions to Assess Outcomes and Disparities in Chronic Kidney Disease among Veterans

Co-Principal Investigator:
Michael A. Langston, Department of Electrical Engineering and Computer Science, University of Tennessee
Abstract:
Chronic kidney disease (CKD) has emerged as a leading cause of morbidity and mortality worldwide, with African Americans having substantially higher incidence rates than whites. Traditional risk factors for cardiovascular disease (e.g., hypertension, hyperlipidemia and obesity) have been proposed as related variables, but they can explain neither the high disease burden nor the excessive racial disparities seen in CKD. Obesity, for example, has even been associated with better outcomes in dialysis patients. Our current CKD understanding is based on scant experimental evidence. Little progress has been made in identifying patient subgroups or demographics that might point to differential diagnosis. Treatments based on complex metabolic abnormalities are limited to short term biochemical endpoints such as phosphorus, hemoglobin and PTH that provide little in the way of convincing evidence of their efficacy in impacting long term clinical outcomes. And treatments aimed at retarding or preventing the development of de-novo CKD and its progression remain limited with little if any recent progress. We therefore seek to leverage the wealth of information contained in large database repositories so that we can systematically examine and compare both traditional and non-traditional risk factors, and also study treatment modalities found in both patients at risk for CKD and in those with established CKD, in particular across races. The national Veterans Affairs research database in particular enables us to perform population-level analysis of multiple disease processes and their treatments. At the same time, it offers us the opportunity to account for complex interactions and to extract granular race/ethnicity and patient-level information. With this wealth of data we will examine, for example, the effect of interventions on clinical outcomes in millions of veterans with non-dialysis dependent CKD, including over a quarter million African Americans. We will also employ powerful graph theoretical algorithms and leadership-class computing resources to test and extend the validity of preliminary findings, to identify putative patient subtypes, to highlight latent, subtle and often highly complex interactions between the various concomitantly administered treatment modalities and clinical outcomes, and to pinpoint outliers that may otherwise confound the analysis.
Research Partners:
The PI for this project is Csaba P. Kovesdy at The University of Tennessee Health Science Center.